Adherence to weight reduction diets has traditionally been viewed as a health-promoting behavior, but little research has been done to investigate the potentially adverse consequences of dieting. The proposed research will employ epidemiologic techniques to explore the health implications of weight change, including weight loss, weight gain, and single or repeated cycles of weight loss and regain. To achieve this aim, we have obtained access to existing data bases from three longitudinal studies, in order to examine the relationships between changes in body weight and various chronic disease outcomes. The endpoints of major interest are all-cause mortality, coronary heart disease, hypertension, all cardiovascular disease, cancer, and diabetes; weight change will also be evaluated in relation to some major risk factors for cardiovascular disease (e.g., blood pressure, cholesterol). Two general types of weight change phenomena will be examined: systematic change over time, and random or periodic fluctuation. Computationally, these are described by compartmentalizing intraperson variability in body weight into linear trend and residual variability. Due to the large amount of previous research on the health consequences of obesity per se, the proposed analysis will control for obesity whenever possible, in order to identify independent effects of weight change. Most of the analyses will be based on data from the Framingham Heart Study and the Gothenburg Prospective Studies of Men and Women. An ancillary project will address a methodologic issue related to validity of biennial measurements, using the Multiple Risk Factor Intervention Trial data base.